1. Magnetoencephalography (MEG) and its role
in studying human neurophysiology
David Poeppel
Cognitive Neuroscience of Language Lab
Department of Linguistics and Department of Biology
Neuroscience and Cognitive Science Program
University of Maryland College Park
Additional slides courtesy of:
• Kanazawa Institute of Technology/Eagle Technology
• Prof. Dr. Kensuke Sekihara, Tokyo
• Prof. Dr. Timothy Roberts, Toronto
• Prof. Dr. Riitta Salmelin, Helsinki

2. resonance imaging (fMRI)
Positron emission
tomography
(PET)
Excellent spatial
resolution (~1-2mm)
Limited temporal
resolution (~1sec)
Hemodynamic
techniques
Electro-magnetic
Functional magnetic
resonance imaging (fMRI)
Non-invasive
recording from
human brain
(Functional
brain imaging)
Electro-
encephalography
(EEG)
Limited spatial
resolution (~1cm)
Excellent temporal
resolution (<1msec)
Magneto-
encephalography
(MEG)
D. Poeppel , A. Braun et al.

3. Origin of the signal - noninvasive measurement - direct measurement.
skull
CSF
tissue
MEG
EEG B
orientation
of magnetic
field
recording
surface
scalp
current
flow
- noninvasive measurement
- direct measurement.

4. requires sensitive detectors (low noise-high gain amplification)
How small is the signal?
Earth field
Intensity of magnetic signal(T)
EYE (retina)
Steady activity
Evoked activity
LUNGS
Magnetic contaminants
LIVER
Iron stores
FETUS
Cardiogram
LIMBS
Steady ionic current
BRAIN (neurons)
Spontaneous activity
Evoked by sensory stimulation
SPINAL COLUMN (neurons)
HEART
Cardiogram (muscle)
Timing signals (His Purkinje system)
GI TRACK
Stimulus response
Magnetic contaminations
MUSCLE
Under tension
Urban noise
Contamination at lung
Heart QRS
Biomagnetism
Fetal heart
Muscle
Spontaneous signal
(a-wave)
requires sensitive detectors (low noise-high gain amplification)
Signal from retina
Evoked signal
Intrinsic noise of SQUID

5. Superconducting Quantum Interference Devices (SQUIDS) with differential measurement
Magnetometer
Gradiometer
KIT System
CTF System
BTi-4D
NeuroMag VectorView
50 mm base line
Planar type
NeuroMag VectorView
BTi-4D Magnes
Axial type

6. Superconductivity To construct a highly sensitive detector
- Magnetic flux quantization
- Josephson effect
- Linearization

7. Capturing the signal For a gradiometer of this
type, a signal from cortex
looks different between the two
coils because of the distance
between the two coils.
A signal from far away, however,
will look similar in size to the
two coils.
This gradiometer principle can
help further with the challenging
problem of measuring small
source that exist in an electro-
magnetic environment with many
large source (subways, elevators,
computers, etc.).
axial gradiometer
recording surface
neuronal source

8. In addition to using gradiometers: Noise reduction using reference channels

9. In addition to using gradiometers and reference channels for noise reduction: Magnetically shielded room (MSR)

12. High density sensor array
front view
bottom view

14. Sensor layout: recording from 160 channels
Response peak at 98ms after onset of an auditory stimulus,
in the left and right temporal lobes.

15. Butterfly plot: overlay of the channels over right temporal lobe
Response peak at 98ms after onset of an auditory stimulus

16. Contour plot: distribution of magnetic field at
peak response

17. For better source reconstruction …
not so
ideal
expensive-
but closer to
ideal
…high spatial sampling is crucial.

18. Magnetic source imaging (MSI): MEG + MRI
Dipole fit at response peak, 98ms after onset of stimulus

19. Somatosensory evoked field (SEF)

20. Dipole locations subsequent to somatosensory and auditory stimulation
SI 40ms
SII 160 ms
and >300ms
AI 100ms
Dipole locations subsequent to somatosensory and auditory stimulation
(primary and secondary somatosensory as well as primary auditory areas
and the time of response peak).
Disbrow et al. (2001) J. Neurophysiol.

21. What is the benefit of using MEG?
EEG
As high temporal resolution as EEG …
… but much easier and quicker to set up (kids, patients)
- Sensitivity to within-subjects effects
Magnetic fields are not differentially attenuated ….
…. easier to get a reasonable estimate of source over time

22. fMRI (yellow blobs) and MEG (red dots)
show remarkably consistent co-localization
Roberts & Poeppel, forthcoming

23. Sanders, Sekihara, Poeppel 2003

24. MEG is a technique that allows you to
(i) record brain activity directly, with
excellent temporal resolution (ms)
(ii) design within-subjects experiments and
evaluate single-subject data
(iii) test models of cognitive processes and
evaluate how these models map on to the brain.